The correct answer is C because the company is not asking for a generic quality score; it needs evaluation against a proprietary de-escalation framework and empathy standard. Salesforce Testing Center supports LLM-judge style evaluations where a judge model compares the agent response against task criteria. For a custom rubric, the specialist should create a custom evaluation and provide a tailored prompt that defines the exact framework criteria. Option A is wrong because standard Coherence is not intended to be rewritten into a proprietary policy rubric. Option B is also too generic because default instruction adherence does not necessarily understand UC’s internal de-escalation model. Custom evaluation is the right mechanism for organization-specific qualitative standards.

In Salesforce Data Cloud, a custom search index is created to enable efficient retrieval of data (e.g., documents, records) for AI-driven processes, such as grounding Agentforce responses. Let’s evaluate the options based on Data Cloud’s functionality.

Option A: A retriever that shares the name of the custom search index.When a custom search index is created in Data Cloud, a corresponding retriever is automatically generated with the same name as the index. This retriever leverages the index to perform contextual searches (e.g., vector-based lookups) and fetch relevant data for AI applications, such as Agentforce prompt templates. The retriever is tied to the indexed data and is ready to use without additional configuration, aligning with Data Cloud’s streamlined approach to AI integration. This is explicitly documented in Salesforce resources and is the correct answer.

Option B: A dynamic retriever to allow runtime selection of retriever parameters without manual configuration.While dynamic behavior sounds appealing, there’s no concept of a " dynamic retriever " in Data Cloud that adjusts parameters at runtime without configuration. Retrievers are tied to specific indexes and operate based on predefined settings established during index creation. This option is not supported by official documentation and is incorrect.

Option C: A predefined Apex retriever class that can be edited by a developer to meet specific needs.Data Cloud does not generate Apex classes for retrievers. Retrievers are managed within the Data Cloud platform as part of its native AI retrieval system, not as customizable Apex code. While developers can extend functionality via Apex for other purposes, this is not an automatic outcome of creating a search index, making this option incorrect.

Why Option A is Correct:

The automatic creation of a retriever named after the custom search index is a core feature of Data Cloud’s search and retrieval system. It ensures seamless integration with AI tools like Agentforce by providing a ready-to-use mechanism for data retrieval, as confirmed in official documentation.

The correct answer is A because chunking divides large source documents into smaller retrievable units, often described as passages or chunks. This is fundamental to retrieval-augmented generation because the retriever should return only the most relevant pieces of content, not entire documents. After chunking, the content can be indexed and vectorized so semantic search can match user questions to relevant passages. Option B is wrong because sending entire documents to the model is inefficient, increases token usage, and reduces retrieval precision. Option C is wrong because chunking strategy directly affects retrieval quality; poor chunk boundaries can omit context or return noisy results. Effective chunking improves grounding accuracy by allowing the agent to retrieve focused, meaningful document segments.

The correct answer is A. Agent-to-Agent (A2A) Protocol is the correct open standard when independent AI agents need to discover each other, communicate securely, exchange structured information, and delegate work across platforms. In this scenario, UC’s agent is not merely retrieving data from an external system; it must communicate with another autonomous logistics agent and delegate execution of a shipping reroute. MCP is used to expose external tools, data, and context to agents, but it is not the primary protocol for agent-to-agent negotiation. OpenAPI describes REST APIs and is useful for traditional service integration, but it does not provide an agent collaboration protocol. Salesforce describes A2A as an open standard for secure communication, task delegation, and collaboration between agents from different platforms.

To address the ineffectiveness of a prompt template reported by a customer, the Salesforce Agentforce Specialist should use the Prompt Builder Scorecard (Option B). This tool is explicitly designed to evaluate and monitor prompt templates against key criteria such as relevance, accuracy, safety, and grounding. By leveraging the scorecard, the specialist can systematically identify weaknesses in the template and make data-driven refinements. While monitoring and refining based on user feedback (Option A) is a general best practice, the Prompt Builder Scorecard is Salesforce’s recommended tool for structured evaluation, aligning with documented processes for maintaining prompt effectiveness. Changing the grounding object (Option C) without proper evaluation is reactive and does not address the root cause.

Salesforce Einstein Agentforce Specialist Certification Guide: Emphasizes using the Prompt Builder Scorecard to evaluate prompts and iterate based on results.

Trailhead Module: " Einstein for Developers " highlights the scorecard as a critical tool for assessing prompt performance.

Salesforce Help Documentation: Details the Scorecard’s role in evaluating prompts against predefined criteria.

Salesforce Agent is designed to provide a conversational AI interface that can be utilized by different types of Salesforce users, such as developers, sales agents, and retailers. It acts as an AI-powered assistant that facilitates natural interactions with the system, enabling users to perform tasks and access data easily. This includes tasks like pulling reports, updating records, and generating personalized responses in real time.

Option A is correct because Agent brings a conversational interface that caters to a wide range of users.

Option B and Option C are more focused on developing and training AI models, which are not the primary functions of Agent.

Salesforce Agent Overview: https://help.salesforce.com/s/articleView?id=einstein_copilot_overview.htm

The Prompt Builder Best Practices Guide emphasizes including a role instruction in prompts to set tone and communication style. The document explains: “Role-playing instructions (e.g., ‘You are an empathetic customer support agent’) are a best practice for ensuring the LLM adopts the desired persona, tone, and style in responses.”

Option A directly reflects this principle by instructing the LLM to role-play as a character, resulting in empathetic, customer-centric responses.

Option B (listing keywords) provides no tone or intent guidance.

Option C (including full interaction history) risks token overload and redundancy, which can degrade prompt performance.

Thus, Option A aligns with Salesforce’s prompt design best practices for tone and empathy.

References (AgentForce Documents / Study Guide):

Salesforce Prompt Builder Design Guide: “Using Role Instructions for Empathy and Tone”

AgentForce Prompt Engineering Handbook: “Persona and Context in Prompts”

AgentForce Study Guide: “Creating Effective Prompts for Customer-Facing Use Cases”

According to the AgentForce Retrieval Optimization Guide, when users ask both structured (exact) and unstructured (natural language) questions, the best practice is to use hybrid search. The documentation states:

“Hybrid search combines the precision of keyword retrieval for structured terms, such as IDs or model numbers, with the semantic flexibility of vector search for natural language queries. This approach ensures both deterministic accuracy and contextual understanding.”

Option A (keyword search only) fails for natural language queries, which require semantic understanding.

Option B (semantic search only) may misinterpret or overlook structured identifiers like product or model numbers.

Therefore, Option C—hybrid search—provides the ideal balance between exact match precision and contextual recall.

References (AgentForce Documents / Study Guide):

AgentForce Retrieval and Indexing Guide: “Choosing Between Keyword, Semantic, and Hybrid Search”

AgentForce Data Cloud Handbook: “Optimizing Multi-Modal Retrieval Strategies”

AgentForce Study Guide: “Hybrid Search for Structured and Unstructured Queries”

Per the AgentForce Security and Permission Management Guide, the AgentForceServiceAgentUserPsg (Permission Set Group) controls access privileges for Service Agents, including which Salesforce objects, fields, and related data they can read or interact with.

When extending an agent’s access to additional custom fields or related objects, the documented best practice is to update the existing permission set group assigned to that agent type rather than creating new or profile-based permissions. This approach maintains centralized permission governance, ensures license alignment, and avoids conflicts or redundancy across multiple permission layers.

Option B is incorrect because creating a new permission set with an Einstein Agent License is unnecessary — the permission set group already includes license mappings. Option C is not recommended, as editing the Einstein Agent User Profile can cause system-wide effects and deviate from Salesforce’s principle of least privilege.

Therefore, the correct approach is Option A – Update the Object and Field access in the AgentForceServiceAgentUserPsg permission set group, ensuring secure and proper access for the Service Agent to custom case data.

Universal Containers (UC) wants to enable its sales team to leverage AI to recommend products from its catalog. The best option for this use case is a Flex prompt template.

A Flex prompt template is designed to provide flexible, customizable AI-driven recommendations or responses based on specific data points, such as product information, customer needs, or sales history. This template type allows the AI to consider various inputs and parameters, making it ideal for generating product recommendations dynamically.

In contrast:

A Record summary prompt template (Option A) is used to summarize data related to a specific record, such as generating a quick summary of a sales opportunity or account, but not for recommending products.

An Email generation prompt template (Option B) is tailored for crafting email content and is not suitable for suggesting products based on a catalog.

Given the need for dynamic recommendations that pull from a product catalog and potentially other sales data, the Flex prompt template is the correct approach.

Salesforce References:

Salesforce Prompt Templates Overview: https://help.salesforce.com/s/articleView?id=000391407 & type=1

Flex Prompt Template Usage: https://developer.salesforce.com/docs/atlas.en-us.salesforce_ai.meta/salesforce_ai/prompt_flex_template